Hi

I have very high CPU usage on my RB750 which causes kernel failure reboots, tried various routerOS versions

If i do profile running i get the following

So you are writing to the flash drive a lot causing the CPU to go up. Are you running a proxy cache? Are you storing logs locally? Are you using user manager?

No not running any web proxy , also switched off all local logging and not running user manager yet this still happens

exebug,
Do you get the issue at RouterOS 5.18?
Please contact MikroTik support (support@mikrotik.com) with attached support output file, when flash process usage is high.

Yes on 5.18

Mailed the supout.rif file

I have the same issue whin a RB750UP.
Flash is using up CPU time causing reboots.
This happens during the day when traffic
at night the routerboard runs smoothly.
Wny solution to this yet?
I cannot even create the Support_Output_File
since flash consumes any CPU time left and
reboots.

I've got the same problem with an RB751U. I turned off web cache, I never had userman on, it's still off. I turned off any local logging. I do still have DNS cache on. The problem started for me when I was testing the web proxy with a USB drive. I saw flash running away, reverted my changes, but ever since this problem has persisted. Can't get supout as previous poster said.

Same problem for me (and a friend) with RB750s in all last releases (in 5.19 too).
When doing a download at about 50Mbit/s flash goes to 100% and winbox will disconnect and the router reboots (if the download takes long enough).

I stopped using it just lots of headache for me.

Switched to 750gl, cpu is still going to 100% but not as often though this flash issues needs to
get addressed.

Sent from my GT-P7510 using Tapatalk

We are seeing this issue on RB750GL, RB750UP and RB2011.

We have logged two tickets with Mikrotik Ticket#2012081366000093 and Ticket#2012041966000517. We were informed the issue is resolved from 5.14, but it certainly is not. We are still experiencing the issue on 5.19.

We need to get this issue resolved, and want to work with Mikrotik to do so, but responses from support@mikrotik.com are extremely slow.

Can you please look if same SAMSUNG K9F1208U0C NAND-Flash are on all these Routerboards?

We are seeing this issue on RB750GL, RB750UP and RB2011.

Nope, mine has ST NAND512W3A2CN6

A friend of mine says the problem disappears as soon as he has Queues affecting the traffic and starts again when he removes them.

My observation is that it only happens on RB750s that also had the bad caps problem.

I had a 750gl lockup on me again. This time its the ram going to
Above 90%. Just wondering if it takes a mass protest to get things
working fine.

Sent from my GT-P7510 using Tapatalk

Try carefully open and look for bulged green capacitors.

I had a 750gl lockup on me again. This time its the ram going to
Above 90%. Just wondering if it takes a mass protest to get things
working fine.

hello.
I also have problems with this router.
Had version 5.20 (today upgraded to 5.21 and bootloader to v 3.0 we will see now)
It locks up almost every day in evening hours when trafic is high.

At first it just slows login procedure and then it stops open certain web pages and soon login becomes imposible for manual reboot not even over telnet.
then internet slows down more to nothing after about half hour after that watchdog reboots router. (thank god)
i looked at logs there is nothing.
We setup reboot of router in early morning for refreshing ram but that did not help.

wierd i have another same router same version on another interent connection and there is not any problems.
I have some indications that there is some very specific trafic to cause this lockup.
Sry for my english.

Hello. Now the other one have same simptoms blocks after aprox 24 hours and is rebooted by watchdog.


I also have new info that helped deblock my router and could log in again over winbox.
I have another routerboard conected over lan to rb750 and when i cant connect to rb750 i connect to closest one and try to go over telnet.
In some cases i manage to enter and do reboot.
Today i did not manage to connect over telnet and i was out of option and then i tryed different things like disabling AP interface to disconect all users but with no luck
Then i disabled LAN port which goes to that RB750 and it worked.
Routerboard normaly operated like it was rebooted. ????
Like there is some problem with lan port or driver that controls it.
Hope it helps someone

tarcius... Have you found a solution to this problem? I have the exact same issue on my routers.

Its just unbelievable the way this isssue isn't addressed. Does it mean I stop
using the routerboard and gove the edgemax router a try.

Its just unbelievable the way this isssue isn't addressed. Does it mean I stop
using the routerboard and gove the edgemax router a try.

In an industry where time is money, it was more expensive for us to follow this through with Mikrotik support, than it was to replace all the mips-be devices with PPC ones. So we did, we replaced all the RB7xx's we had this issue on with RB1200's and they are fine.

For us it appeared to be triggered by SNMP on these devices, the flash and management processes would spike up and consume the CPU. We also noticed that it only happened on devices where high throughput/PPS were going through the bridge/routing stages, on devices that were just forwarding packets with MPLS we have never had an issue, however on devices doing L3 routing, or bridging a VPLS tunnel to a interface the issue would always occur.

Its just unbelievable the way this isssue isn't addressed. Does it mean I stop
using the routerboard and gove the edgemax router a try.

In an industry where time is money, it was more expensive for us to follow this through with Mikrotik support, than it was to replace all the mips-be devices with PPC ones. So we did, we replaced all the RB7xx's we had this issue on with RB1200's and they are fine.

For us it appeared to be triggered by SNMP on these devices, the flash and management processes would spike up and consume the CPU. We also noticed that it only happened on devices where high throughput/PPS were going through the bridge/routing stages, on devices that were just forwarding packets with MPLS we have never had an issue, however on devices doing L3 routing, or bridging a VPLS tunnel to a interface the issue would always occur.

I noticed the same issue. I had a RB750UP at a tower working as a PPPoE server, hotspot and basic routing it could lockup every single day with the CPU at 100%. I moved to the Internet entry point just doing Natting and load balancing and It has never locked up since. We have started replacing the routers one at a time.

Seems a overkill to replace a rb750 with a with a rb1200. i just need pppoe, ospf and hotspot. For now i set all my routers to reboot every night which seems to help a bit. I have one router though that keeps locking up every morning even if i reboot it. Could multiple dude servers have an effect? Perhaps this could be the reason why the cpu runs up. Also found a thousand api logins from my radius server. Sent a support rif and they reckon i should change the routers as they don't have enough memory. Upgraded firmware to 5.22 with same results, in fact it was worst. Downgraded to to 4.17 to get stability again.

Ever a resolution on this item? Just now seeing this and dont want to dig through changelog.

Read through the changelog. Doesnt look like it. Cool. Guess ill be replacing it with a 450G

We've experienced this issue with several RB750's. Processor at 100%, Winbox unable to display information in Interface and Queue fields among others. Profile indicates Flash is consuming all resources. CPU typically at less than 20% until the bug surfaces. Seems like a simple fix once the cause is identified. I just let them run and they eventually recover. Without knowing the filesystem and other parameters, any suggestion is conjecture. I'd like to see it addressed as it is a reproducible behavior. But, we ask a lot of these SOHO routers so an upgrade to the RB1100 is realistic.

Same problem here. It seems happens when plug/unplug an eth cable, and there are in total more than 2 devices connected

Inviato dal mio M-MP875S2 con Tapatalk now Free